Synthesis of nanocrystalline titanium nitride coatings from the plasma of a composite-cathode arc discharge Yu. F. Ivanov, N. N. Koval, and O. V. Krysina Institute of High Current Electronics, SB RAS, 2/3 Akademichesky Avenue, Tomsk, 634055, Russia ∗ T. Baumbach, S. Doyle, and T. Slobodskyy Karlsruhe Institute of Technology (KIT), Institute for synchrotron radiation, Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-Leopoldshafen, Germany, EU 1 1 N. A. Timchenko, R. M. Galimov, and I. P. Chernov 0 Tomsk Polytechnic University, Tomsk, 634050, Russia 2 n A. N. Shmakov a Siberian synchrotron radiation centre, Budker Institute of Nuclear Physics, J SB RAS, Lavrentyev av. 11, Budker INP, Novosibirsk, 630090, Russia 0 1 Experimentaldataaregivenonthestructureandpropertiesofnanocrystallinehardeningcoatings of titanium nitride doped with copper, produced by plasma-assisted vacuum arc deposition by ] evaporating powder cathodes. A model of nanostructurization of this type of coatings is proposed. i c Accordingtothemodel,thenanocrystallizationinthesematerialsisduetothedopantatoms,which s form an amorphous sheath around a crystallite, thusdefiningits size. - l r mt I. INTRODUCTION ponent arc-dischargeplasma for the formation of nanos- tructured nitride coatings, such as: . t (1) production of several metal plasma flows (several a The vacuum-arc method of coating deposition based m on the generation of highly ionized metal plasma flows single-element cathodes) in the presence of nitrogen re- active gas and - by an arc discharge is widely used to synthesize thin d functional coatings1,2. The coatings are formed due to (2) evaporation of composite cathodes whose mate- n condensation of the plasma of the eroding cathode ma- rial includes several elements (mosaic cathodes, sintered o terial on the substrate surface. The cathode material powdercathodes,cathodesproducedbyself-propagating c [ can be in fact any metal, alloy or metal-base compos- high-temperature synthesis). ite. With a reactive gas present in the discharge gap, Each of these methods has some advantages and dis- 1 a layer based on the compounds of the cathode mate- advantages and, undoubtedly, demands optimization of v rial and working gas elements (nitrides, oxides, and car- the parameters of the coating deposition process. In our 4 5 bides)issynthesized. Thehighdegreeofionizationofthe previous experiments11–13 we have revealed that Ti-Cu, 7 vacuum-arc plasma and the possibility of controlling the Ti-Si, andTi-Al composite cathodes producedby sinter- 1 coating synthesis parameters (working gas pressure, dis- ing metal powders in vacuum14 are best suited for the 1. charge current, bias voltage, etc.) over wide limits make synthesis of superhard nitride coatings by the vacuum- 0 it possible to produce a desired effect on the structural, arc deposition technique. 1 physical, and mechanical characteristics of the conden- Alongside with extensive studies on the plasma- 1 sate. A promising application of the vacuum-arccoating assisted synthesis of nanocrystalline coatings, investiga- : v deposition technique is the production of coatings with tionsoftheirstructure-phasecomposition,propertiesand i crystalliteslessthan100nminsize. Coatingsofthistype of the effect of the doping element on the way by which X featuresuperhardness( 40GPa),highresistancetowear nanostructurization of the coating occurred are carried ar and oxidation, etc. Exp≥eriments on the plasma-assisted out. Twobasicmodelsofthestructureformationinthese vacuum-arccoatingdepositionhaveshownthatthegrain coatingsdependingontheirphasecompositionhavebeen sizecanbedecreasedbyaddingsmallamountsofdoping proposed. A detailed description of the model classifica- elements (Cu, C, Si, Al), whichrestrictthe graingrowth tion is given elsewhere15. In the first case, a coating inthecoatingbasematerialduringthecoalescenceofnu- of the nc-MeN/hard phase type is formed with a-Si N , 3 4 cleation centers, to the coating composition3,4. Results BN, etc. playing the part of the hard phase. For in- ofthe pioneeringstudies onthis subjectcarriedoutwith stance, in Ti-Si-N coatings, nanosized crystallites of the the use of CVD processes5 and magnetron sputtering6 basic phase (TiN) are embeded in an amorphous silicon were published in the late 90-s. Now extensive investi- nitride (Si N ) matrix. In the second case, when metals 3 4 gations on the production of multicomponent coatings, whichdonotformcompoundswithnitrogen,suchasCu, such as Ti-Si-N, Ti-Al-N, Zr-Cu-N, and Ti-Si-Al-N, are Ni,Y,Au,etc.,areaddedtoMeNcoatings,atomsofthe under way7–10. doped element surroundcrystallites of the basic element Therearesomemethodsofthegenerationofmulticom- nitrides, thus restricting their growth on the nanometer 2 scale,andacoatingofthenc-MeN/softphasesystem(Me 50 Ti-Cu-N = Ti, Zr, Cr, Ta, etc.) is formed. Unfortunately, the ca- Ti-N pabilities of the existing research techniques restrict the ) scope of investigation of nanostructured and amorphous mN40 materials. Therefore, direct verification of the proposed d ( models of nanostructurization of multicomponent coat- a30 o ings with a certain coating is not always possible, and al l the use of complex procedures, such as those based on m20 the properties of synchrotron radiation, is required. or N This paper presents a study on the structure, phase 10 composition,andelementcompositionofnitridecoatings synthesized on metal substrates by evaporating Ti cath- 0 odeandTi-Cucompositecathode. Thegoalofthestudy 0 50 100 150 200 250 300 Indentation depth (nm) was to elucidate the effect of the doping elements on the featuresoftheformationofnanocrystallinestructure,on FIG. 1. Loading-unloading curves for coatings produced by the basic phase crystallite size, and on the properties of evaporatingTi andTi-12 at. % Cu cathodesin arc-discharge the coatings. plasma. II. THE TEST MATERIAL AND THE ness was 3-5 µm at a coating growth rate of 1-3 µm/h. EXPERIMENTAL PROCEDURE Investigations of the deposited coatings were car- ried out by the following methods: optical microscopy The deposition of nitride coatings in low-pressure (OLYMPUS GX71), transmission (EM-125) and scan- arc discharges was carried out on a plasma-ion set-up ning electron microscopy (Philips SEM 515 equipped equipped with a standard arc evaporator and PINK, an with EDAX ECON IV, an element composition micro- original gas-plasma generator developed at the Institute analyzer), micro- and nanoindentation (PMT-3, NHT- of High Current Electronics, SB RAS (Tomsk)16. Addi- S-AX-000X Nano Hardness Tester), scratch testing tionalionizationoftheworkinggasbymeansofaplasma (MST-S-AX-000Micro-ScratchTester),andtheCalotest sourcewithfilamentcathodemadeitpossibletoincrease method. Thephaseandelementcompositionwasinvesti- theefficiencyofpreliminarycleaningofthespecimensur- gatedon powder diffraction stations, on the synchrotron face and to realize the formation of nitride compounds radiation (SR) channels of the VEPP-3 (RFA-SR, PRD under the conditions of plasma assistance. SR) energy storage ring and on the PDIFF beamline of The generation of multicomponent plasma and the the ANKA SR source. subsequent condensation of coatings were carried out by evaporating cathodes of the compositions Ti containing III. RESULTS AND DISCUSSION 12 at. % Cu in a nitrogen medium. To compare the coatings obtained and the widely used two-component coatings by their structure-phase and element compo- Measurements of the micro- and nanohardness of the sition and by mechanical properties, conventional TiN test coatings were performed in order to investigate the coatings were produced and investigated. Beryllium foil influence of crystallite size on mechanical properties of ofthickness0.5mm,WC-8%Cohardalloy,andAISI304 the samples. The measurements performed at a normal steelwereusedassubstratesforX-rayanalysis,mechani- load of 500 and 50 mN, respectively, have shown that cal,and TEMinvestigations,correspondingly. After me- the hardness of the coatings produced by evaporating chanicalgrinding,polishing,andwashinginanultrasonic composite cathodes is greater than the hardness of ti- bath, the specimens were placed, on a substrate holder, tanium nitride ( 25 GPa) by a factor of 1.5-2. The ≈ ≈ in a vacuum chamber at a distance of 300 mm from the coatings obtained with the use of Ti-12 at. % Cu cath- cathode. Immediately before the deposition of coatings, odes possess superhardness ( 40 GPa). Analysis of the ≈ the surface of specimens was cleaned and activated by loading-unloadingcurves(Fig.1)obtainedbythenanoin- bombardment with accelerated argon ions at a negative dentation method has given the elastic strain of the test substrate potential of 1 kV. During the bombardment, coatings. The greatest residual strain of 75% was ob- ≈ ◦ the specimen surface heated up to 300 C. servedforatitaniumnitridecoatingandtheleastof50% To improve the adhesive characteristics of a nitride for Ti-Cu-N coatings. The degree of elastic recovery of coating, its formation was preceded by deposition of a thesurfaceshapeforthe coatingsformedbyevaporating sublayer of thickness about 100 nm by evaporating the powder cathodes was 2 times greater than that for the cathode material in an argon medium. Synthesis of all TiN coatings. The Young modulus of the multicompo- multicomponentcoatingswascarriedoutinthefollowing nent coatings was in the range 500-550GPa. parameterranges: Ub = (100 300)V,p=0.3 0.4Pa, The scratch-test method was used to determine the − − ◦ − Id =50 100A,andT =300 400 C.Thecoatingthick- critical load at which destruction of a coating begins. − − 3 a) b) 0.20 N TiN s) ) TieO N TiN TiCuN b. unit0.15 (111100) B eO200) Ti Be Be (220) Be ntensity (ar0.10 (101) TiN0.3(002) B(101) BeO (311) TiN 22) TiN I ( (e 2 0.05 B ( 25 30 35 40 45 50 55 60 100 nm 2theta (degrees) FIG. 3. Diffraction patterns of TiN coatings without doping (TiN)anddopedwithCu(Ti-Cu-N)obtainedforthephoton c) energy of 10.5 keV properties and to answer the major questions of which is the role of the impurity atoms in the formation of the nanocrystalline structure of synthesized coatings, where theyarelocalized,andwhethertheyformtheirowncrys- tallographic phase. It is necessary to notice, that earlier the composite cathode material was investigated14. It was showedthat copper in cathode material is situated uniformly mainly on boundary of main phase (α-Ti) or in compound with Ti (CuTi ). The size of titanium based phases is in the 2 range of 3-20 µm for Ti-12at. % Cu cathode. The size FIG. 2. (Top) Image of the structure of a Ti-Cu-N coating formed by evaporating Ti-12 at. % Cu composite cathode in ofcathode spots ofvacuum arcis in rangeof50-300µm. a nitrogen: dark field in the reflection of a type 111 ring Thisfactisevidenceofuniformlyevaporationofcompos- of TiN (a), electron diffraction pattern (b). (Bottom) TiN ite cathode elements. crystallite size distribution in the Ti-Cu-N coating obtained Investigations of hardening layers based on titanium by evaporation of Ti-12at. % Cu cathode (c). nitridewithadditionalelementscrystallinestructureand element composition have been carried out on the sta- tionsofthestorepowderdiffractionandx-rayfluorescent For TiN, this quantity was about 3.6 N, whereas the de- element analysis (VEPP-3)18 andof the PDIFF bamline structionofTi-Cu-Ncoatingsof6.0Nbeginsatacritical ofANKAsynchrotronradiationsource19. Werelyonthe load 2 times greater than that for TiN coatings. data for the constructionof a model of the coatings syn- thesis. The results of these investigations are given in With the use of transmission electron diffraction mi- Fig. 3 and Fig. 4. croscopyof thin foils it has been found that the coatings Figure 3 presents diffraction patterns of TiN coatings formed by evaporation of composite cathodes consist of without admixture and doped with Cu on a beryllium nanosized crystallites oriented randomly relative to each substrate of thickness 0.5 mm. The beryllium substrate other. This follows from the strongly pronounced ring shows up in all diffraction patterns in the form of very structure of the electron diffraction patterns [Fig. 2(b)]. narrow strongly pronounced reflections. The reflections Electron diffraction analysis has shown that the crys- fromthesynthesizedcoatingstructurearewiderowingto tallites of the coatings consist of δ-TiN. The measure- their nanocrystalline structure. As well known for pow- ments of size of coating crystallites were carried out on der diffraction20, the physical widening of reflections, β, dark-lightimages,the averagesizeofcrystallitewasesti- can be used to estimate the crystallite size in the direc- mated by the methods of statistical analysis17. The size tion perpendicular to the reflecting plane with indices of crystallites in the coatings containing copper is equal hkl: to 10-30 nm [Fig. 2(a), Fig. 2(c)]; the coatings produced byevaporatingpuretitaniumhavecrystallitesofaverage size 100 nm. Dhkl =nλ/βcosθ (1) The investigations performed by the above methods gave insufficient information to construct a model of the where D is the size of coherent-scattering region in processes responsible for the variations of the coating angstroms,λistheradiationwavelength,θisthescatter- 4 1.0 100 00.1 6 9 00.2 s)0.8 4. 80 nit Ti nt] 00.4 u e 00.8 rb. 0.6 erc 60 01.6 a p ( 7 [ 03.2 nsity 0.4 1 u 8.9 ping 40 0162..48 e 1 C o Int0.2 e 7. D 20 25.6 F 0.0 0 5 6 7 8 9 10 11 0 5 10 15 20 25 30 Energy (keV) d [nm] FIG.4. X-rayfluorescent spectraof Ti-Cu-Nspecimens near FIG. 5. Simulation of maximum grain size at a given dop- the K-edges of Ti and Cu obtained for the excitation with ing concentration. Thelines correspond todifferent coverage 20 keV photons. needed to prevent further crystallite growth in nm. Our ex- perimental conditions are indicated by a cross. ingangle,βisthephysicallinewideninginthediffraction patterninradians(onthe2θ scale),andnisacoefficient depending on the particle shape and close to unity. R3 x=1 (2) Calculations of crystallite sizes using Eq. 1, where β − (R+dR)3 was computed by approximating the reflection lines by the Gauss method, are in agreement with the crystallite Plot of the grain size dependance on the doping ma- sizes measured by electron microscopy methods. terialconcentrationrequiredtocompletelycoveragrow- ing grain for different cover layer thicknesses shown in The diffraction patterns of the layers of titanium ni- Fig. 5. From the figure it is easy to see that the grain trideandoftitaniumnitridedopedwithcopperareiden- size will grow with decrease in the dopant concentration tical [see Fig. 3]. The location of reflections testifies to until other processes will not start to dominate limiting the presence of TiN, TiN0.3, and Ti phases in the coat- thesizeofthegrain. Inthecaseofgrainsizesbelow5nm ings. Shifts of reflections are not observed in the diffrac- the dependance might be slightly modified by triangula- tionpatterns obtainedfor specimens dopedwith copper. tion correction. Thesedataallowtheconclusionthatcopperatomsdonot Ifthe grainsizesanddopingconcentrationsareknown form compounds with titanium or nitrogenand, equally, the cover layer thickness is expressed as: do not form their own crystalline phase. As the data of x-rayfluorescentanalysispresentedinFig.4 confirmthe presence of copper in the test coating in amounts corre- 1 dR=R( 1) (3) sponding to its content in the composite cathode evap- √31 x − orated in an arc discharge (Ti-12 at. % Cu), it can be − concluded that copper is in the amorphous state at the For our experimental conditions, with 18 nm average crystallite boundaries. No signal from heavy elements grainsizeand12at.%Cuconcentrationwefind0.74nm is present in fluorescence data. An stand alone peak at cover layer thickness (cross in the Fig. 5) corresponding 85 keV is attributed to Be windows. Therefore, we con- to 2.8 single atomic layers coverageof the grain(Atomic cludethattheonlytimeittakesforcopperatomstoform radius of copper is equal to 132 pm). a closed sheath around a growing TiN crystallite deter- mines the time of growth of the crystallite and, hence, its size. IV. CONCLUSIONS In order to explain the observed phenomenon we de- veloped an intuitive model of the growth process. The By method of transmission electron diffraction mi- model assumes that a layer of finite thickness should be croscopyofthinfoilstheformationofthenanocrystalline formedaroundagrowinggraininordertopreventfurther coatingsproducedbyevaporatingTi-Cucompositecath- growthofthegrain. Sphericalformofthegrainischosen odes in low-pressure arc discharges in the presence of as a statistical approximation. For a given grain radius nitrogen plasma was revealed. The coating crystallites R the volume of the grain is calculated as V = 4πR3 consistsofδ-TiN.Forthecopper-containingcoatingsthe 3 similarly the volume of a cover layer with thickness dR averagecrystallitesizeis10-30nm;thecoatingsproduced is calculated as dV = 4π((R+dR)3 R3). The concen- byevaporatingpuretitaniumhavecrystallitesofaverage 3 − tration of doping material can be deduced as: size 100 nm. 5 Withthepowderdiffractionmethodusingsynchrotron formed around of TiN crystallites, preventing their fur- radiation it has been found that copper does not partic- ther growth. ipate in the formation of crystal phases in the coating; that is, they are in the amorphous state. It has been revealed that the copper concentration in the Ti-Cu-N ACKNOWLEDGMENTS coatingssynthesizedbythevacuum-arcmethodcoincides with that in the evaporated cathode (12 at. %). The dataobtainedconfirmthe modelofnanostructur- The work was partly supported by SB RAS under In- ization of coatings based on titanium nitride, according tegrationProjectNo. 43,byRFBRunderGrantNo. 08- towhichthenanocrystallizationofthesematerialsoccurs 08-92207-NSFC-a,andby PresidiumRASunder Project as an amorphous sheath of doping elements (copper) is No. PP27/09. ∗ [email protected] 13 I.M. Goncharenko, Yu.F. Ivanov, M.I. Lobach, O.V. 1 V.A. Barvinok, V.I. Bogdanovich. Physical Grounds Krysina,G.A.Pribytkov,I.A.Andreeva,V.V.Korjova,9th and Mathematical Simulation of Ion-Plasma Evaporation InternationalConferenceonModificationofMaterialswith (Mashinostroenie Press, Russia, Moscow, 1999). Particle Beams and Plasma Flows: Proceedings. Tomsk: 2 A.A. Andreev, L.P. Sablev, V.M. Shulaev,S.N. Grigoriev. Publishing house of the IAO SB RAS, 21-26 September, Vacuum-arc devices and coatings. (NSC KIPT Press, 430 (2008). Ukraine, Kharkov,2005). 14 G.A. Pribitkov, E.N. Korosteleva, S.G. Psakhie, I.M. 3 J. Musil and P. Zeman, Solid State Phenomena 127, 31 Goncharenko, Yu.F. Ivanov, N.N. Koval, P.M. Schanin, (2007). A.V. Gurskih, V.V. Korjova, Yu.P. Mironov, 7th Interna- 4 S. Veprek, M.G.J. Veprek-Heijman, P. Karvankova, J. tional Conference on modification of materials with par- Prochazka, Thin Solid Films 476, 1 (2005). ticle beams and plasma flows: Proceedings. Tomsk: Pub- 5 S. Veprek,S.Reiprich, Thin Solid Films 268, 64 (1995). lishing houseof theIAO SBRAS,163 (2004). 6 J. Musil, J. Vlcek, Thin Solid Films 343-344, 47 (1999). 15 J. Musil, P. Baroch, P. Zeman, (p.1, Research Signpost 7 S. Veprek, M.G.J. Veprek-Heijman, R. Zhang, J. Phys. Publisher, India, 2008). Chem. Solids 68, 1161 (2007). 16 L.G. Vintizenko, S.V. Grigoriev, N.N. Koval, V.S. Tolka- 8 S. PalDey, S.C. Deevi, Mater. Sci. Eng. A 361, 1, (2003). chev, I.V. Lopatin, and P.M. Schanin, Russian Physics 9 P. Zeman, R. Cerstvv, P. H. Mayrhofer, C. Mitterer, J. Journal 9, 927 (2001). Musil, Mater. Sci. Eng. A 289, 189 (2000). 17 Gareth Tomas, Michael J. Goringe, (John Wiley & Sons 10 S. Carvalho, L. Rebouta, A. Cavaleiro, L.A. Rocha, J. Press, New York-Chichester-Brisbane-Toronto, 1979). Gomes, E. Alves, Thin Solid Films 398-399, 391 (2001). 18 B.P. Tolochko, D.I. Kochubey, A.N. Shmakov, K.A. Ten, 11 I.M. Goncharenko, Yu.F. Ivanov, Yu.A. Kolubaeva, K.A. M.R.Sharafutdinov,S.B.Erenburg,Proceedingsofthe5th Koshkin, O.V. Krysina, Vacuum nanotechnologies and International specialized exhibition ”LaboratoryExpo’07” equipment 1, 221 (2006). (NSC KIPT Press, Ukraine, (Expodesign Press, Russia, Moscow, 2007). Kharkov,2006). 19 ANKA - Instrumentation Book (ISS Institute for Syn- 12 N.N. Koval , Yu.F. Ivanov, I.M. Goncharenko, O.V. chrotron Radiation, Forschungzentrum Karlsruhe GmbH, Krysina, Yu.A. Kolubaeva and K.A. Koshkin, Russian member of theHelmholtz Associatio, 2008). Physics Journal 2, 146 (2007). 20 G.LipsonandG.Stiple.InterpretationofPowderPatterns (Russia, Moscow, 1972).